JPH0648239B2 - Corrosion prediction method for buried pipes - Google Patents

Description

TECHNICAL FIELD The present invention relates to a method for predicting corrosion of a buried pipe in soil.

Conventional technology Corrosion of buried pipes is governed by the structure of pipes and environmental factors. Above all, in so-called soil corrosion, which excludes corrosion due to stray current and long-circuit current, the influence of soil corrosivity is relatively large.

Empirical corrosive soils have long been known as coal husks, humus, peat, and submarine mud. In recent years, geological features have been frequently exposed due to the development of hilly areas in the suburbs of the city. Marine groups such as the Osaka group marine clay and the Kazusa group mud layer (called Dotan) are extremely corroded. It turned out that it shows sex. This marine layer is a clay layer deposited on the seabed of a quiet inner bay about 300,000 years ago, and it is now exposed to the hilly areas around the sedimentary plain. Marine clay is characterized by containing a large amount of sulfur compounds such as iron sulfide as compared with fresh water clay and producing sulfuric acid when weathered. It has not been fully clarified how clay having such characteristics is involved in iron corrosion, but
As a result, iron often corrodes heavily in the presence of marine clay.

On the other hand, buried pipes for water, gas, etc. are laid in a wide area, and it is becoming more important to predict the corrosion of pipelines in terms of maintenance. Until now, the external corrosion of buried pipes has been predicted by qualitatively determining environmental factors such as the type of buried soil and the properties of groundwater, and showing the results of soil measurements in ANSI (American standard) and DIN (German standard). It was possible to quantify it to some extent by the method described above.

However, these conventional corrosiveness evaluation methods are useful in considering the anticorrosion design of the newly designed pipeline, but in order to perform rational preventive maintenance, it is more accurate. There is a problem that corrosion prediction is required.

Therefore, an object of the present invention is to solve such a problem and to enable corrosion prediction with higher accuracy than before.

Means for Solving the Problem In order to solve the above problems, the present invention, the growth rate of the pitting depth (P) indicating the degree of outer surface corrosion of the cast iron pipe buried in the soil time (t) power function According to the equation, P = kt ⁿ , and the above (k) is quantified by multiple regression analysis by attribute change assuming that it depends on the environmental factors of the burial site, and then (n) is the measured value of the corrosion depth. This is obtained as a linear model regression coefficient to the difference from the predicted value due to environmental factors.

Effect In this way, the dependence of the corrosion of the cast iron pipe on the environmental factors is first quantitatively obtained by a statistical method. Next, by subtracting the degree of dependence on the environmental factors thus obtained from the measured value of the corrosion depth, the value of the parameter for the burying period is statistically obtained.

Example One example of the present invention will be described below.

First, the survey of soil and buried pipe will be explained. The samples used for analysis were obtained by investigating existing water pipes. That is, the burial conditions were investigated at the excavation site, and the corrosion depth was measured in the range of the pipe length of about 1 m. Also,
The soil around the tube was taken back to the laboratory for analysis. Table 1 shows the survey items and their definitions.

The study area is a residential land development site with a marine clay layer. As shown in Table 2, marine clay has higher salt content and sulfur content than other soils. If the marine clay is forcibly oxidized by the hydrogen peroxide solution, the pH becomes strongly acidic to 1-2. ANSI standard A21.
When the corrosiveness is evaluated according to 5, marine clay has a score of 1
It is 9.5 to 23.5 points, indicating that strong corrosion is exhibited.

Marine clay is a mud originally deposited on the sea floor, and it is thought that it basically satisfies the growth conditions of sulfate-reducing bacteria that cause microbial corrosion. The concentration of sulfate is high,
The reason for this is that the anaerobic conditions are not met in terms of soil quality.

These characteristics of marine clay were used as a reference when conducting the following statistical analysis.

Next, the analysis procedure will be described. The data was analyzed for the purpose of considering the cause of corrosion of the buried pipe and creating a corrosion prediction formula. The analysis was performed by a general statistical method.
The analysis procedure is shown in FIG.

First, for each sample obtained from the survey, it was examined whether or not there is a contradiction in terms of soil corrosion in terms of specific technology. Next, examine the distribution of the data by means of the average value, standard deviation, histogram, etc., and especially for the maximum corrosion index (= maximum corrosion depth of a specific sample / average value of the maximum corrosion depth in the population), which is the objective variable, the probability The distribution was confirmed with paper. Subsequently, the relationship between the two variables was examined by the correlation coefficient and the scatter plot, and the contradiction in the correlation of each variable and the abnormal data were checked again.

In addition, the variables were classified by principal component analysis and used as a reference when selecting the variables to be included in the corrosion prediction formula. The corrosion prediction formula was set to P = kt ⁿ , the buried environment was summarized by the parameter k, and k was determined by multiple regression analysis using attribute variables (so-called forest quantification type I). Furthermore, the parameter n was obtained as a regression coefficient to the time t of the difference between the actual value of the corrosion depth and the predicted value due to environmental factors.

Next, the analysis result will be described.

First, a single correlation analysis was performed.

A correlation coefficient is used to quantitatively express the strength of a linear relationship (correlation) between two variables. It can be judged that the closer the absolute value of the correlation coefficient (r) is to 1, the stronger the linear relationship between the two variables.

Table 3 shows the results of organizing the data obtained so far and calculating the correlation coefficient between each variable.

Here, with respect to the corrosion index (Y ₀ ) indicating the degree of corrosion, variables that showed a strong correlation are H ₂ O ₂ pH, water content ratio, ANSI, cutting and assembling, Redox,
It was ρ ₅ , ρ ₁ , groundwater, soil, sulfide, etc.

From the results of simple correlation analysis, the following tendencies are judged.

(1) The lower the pH value (acidic) after oxidation with H ₂ O _{2, the} deeper the corrosion depth.

(2) The higher the water content, the deeper the corrosion depth.

(3) The larger the evaluation score according to ANSI A21.5, the deeper the corrosion depth.

(4) The embankment has a deeper corrosion depth than the cut soil.

(5) The lower the Redox potential, the deeper the corrosion depth.

(6) The lower the resistivity of soil and extracted water, the deeper the corrosion depth.

(7) Corrosion depth is deeper when there is spring water.

(8) In the soil, the finer the particles, the deeper the corrosion depth.

(9) Corrosion depth is deeper with sulfide.

(10) Soil extracted water having a low resistivity tends to have a low H ₂ O ₂ pH in the soil and a high SO ₄ ²⁻ concentration in the extracted water.

(11) There is a strong correlation between the evaporation residue of extracted water and the concentration of SO ₄ ^2- . (The main component of the solution salt in the extracted water is sulfate.) By the way, the relationship between qualitative characteristics such as cut and embankment, the presence or absence of spring water, and quantitative characteristics such as corrosion depth should be discussed. It is not appropriate to evaluate it by the number of relations. Therefore, the difference between population averages of corrosion depths was tested as follows for the distinction between cut and fill and the presence or absence of spring water.

Statistically, if | Zo |> 1.96, A and B are at a risk rate of 5%.
It is determined that there is a difference in the population mean of. That is, here
It can be said that the embankment is deeper than the cut soil, and the corrosion depth is deeper with spring water.

Next, the classification of environmental factors will be described. Single correlation analysis is 2
Principal component analysis was attempted to clarify the characteristic structure across more variables, while finding the correlation between variables (single correlation). In Fig. 2, the first main component-the second
The position of each variable in the principal component coordinates is shown. Grouping is performed as shown by the broken line in the figure, and each group is shown in Table 5.
Interpreted as having the characteristics shown in.

Next, the quantification of corrosive environmental factors will be explained. The above single correlation analysis and principal component analysis revealed a linear relationship between each variable. As a result, there is no particular abnormality in terms of inherent technology. However, the environmental factors taken up here include some qualitative variables, and even quantitative variables do not always have a linear relationship with the corrosion index. Instead of the usual multiple regression analysis, the so-called Hayashi quantification type I method using dummy variables was used for the analysis.

A regression analysis was performed using the corrosion index as the objective variable. First, the distribution of the corrosion index was investigated. From the plot on the probability paper of FIG. 3, it was judged that the distribution of the maximum corrosion index was in accordance with the logarithmic normal. Therefore, the maximum corrosion index applied the quantification type I after logarithmic conversion.

Some explanatory variables were selected from environmental factors by considering the results of simple correlation analysis and principal component analysis, and adding proper technical judgment. In particular, the following six variables were selected considering the characteristics of marine clay and the interaction between the variables.

By the method of quantification type I, the regression equation of the corrosion index with respect to the variables of ANSI ρ ₁ H ₂ O ₂ pH, other sewage water content ratio, and embankment was obtained. The results are shown in Table 6.

From these results, it was H ₂ O ₂ pH and groundwater that had the greatest influence on the corrosion index, followed by the water content ratio and fill. In other words, if the H ₂ O ₂ pH is less than 3 (it can be judged that it is almost marine clay) and there is groundwater, the water content ratio is 35% or more (it can be judged that the buried soil is clayey or silty and has a large amount of water). ) Showed that the case of the embankment is a highly corrosive environment. Also, when comparing the degree of influence on the corrosion index for each variable alone, H ₂ O ₂ pH, cutting,
The degree of influence is large in the order of ρ ₁ . Therefore, when assessing the burial environment from the aspect of corrosiveness, it is important to check whether it is marine clay or embankment.

Here, the degree of influence of environmental factors on the maximum corrosion depth is Y
It can be quantified by the equation (1) with c as the maximum corrosion index.

Yc = expA ... (1) From Table 6, A = (O ・ δ ₁₁ -0.11 ・ δ ₁₂ ) + (O ・ δ ₂₁ -0.09 ・ δ ₂₂ -0.28 ・ δ ₂₃ ) + (O ・ δ ₃₁ -0.15 ・ δ ₃₂ -0.26 ・ δ ₃₃ -0.49 ・ δ ₃₄ ) + (O ・ δ ₄₁ -0.18 ・ δ ₄₂ -0.32 ・ δ ₄₃ -0.42 ・ δ ₄₄ ) +0.58 (δ = 1 or 0) The method will be described.

Predicting the corrosion depth of buried pipes from environmental factors is (1)
It is possible by formula. However, equation (1) is not practical because it does not consider the speed factor. Here, we tried to consider the burial period.

Generally, in soil corrosion, there is no linear relationship between the amount of corrosion and the burial period. The corrosion rate tends to be high in the initial stage and decrease over time. It is said that the relationship between the maximum pit depth P in the soil and the burial period t is expressed by equation (2) as described above.

P = kt ⁿ (2) Equation (2) can be transformed into Equation (3) by taking the logarithm of both sides.

logP = logk + nlogt (3) Here, the pitting depth P is expediently set to the maximum corrosion index Y ₀ ,
It was considered that k was a parameter indicating the corrosiveness of the environment. That is, Eq. (4) was assumed as a regression model of the corrosion index Y ₀ .

Here, i = 1, 2, ..., M (number of samples) R: number of factors C _j : number of categories in factor j δ _i (jl) = 1 (when solid i reacts to category l of factor j) ) Or 0 (at other _times ) a _jl , a _{R + 1} : coefficient e _i : error term The first term on the right side in equation (4) corresponds to k, which is
It is calculated by the equation (1). Therefore, the parameter n
Can be obtained as a regression coefficient of the linear model by transforming equation (3) into equation (5).

logY ₀ -logYc = C + nlogt (C: constant) (5) That is, an attempt was made to explain the rest of the corrosion index explained by the environmental factors by the burying period. Here, the parameter n
In order to investigate whether or not the value depends on the corrosiveness of the environment, the samples were divided into two groups according to the corrosiveness, and n was calculated for each group. The corrosiveness of the environment was obtained from the equation (1), and the samples were divided by the corrosion index Yc = 1.

The results of the regression analysis are shown in Table 7.

From these results, the regression coefficient n was at a significance level of 1% in each case, and n was a constant value of 0.41 to 0.42 regardless of the corrosiveness of the environment.

Therefore, the corrosion index can be predicted by Eq. (6) depending on the corrosiveness of the buried environment and the buried period.

here, The multiple correlation coefficient R between the actually measured value and the predicted value by the equation (6) was 0.742. Measured value of corrosion index Y ₀ and predicted value The scatter plot of is shown in FIG. In addition, the standardized residuals e _s were plotted against the predicted values in order to confirm the presence or absence of defects and prediction bias in the regression model obtained here. here, Is the residual Is the standard deviation of. As shown in FIG. 5, it is judged that the corrosion prediction model obtained here does not have a peculiar tendency.

As a sample of the survey results of cast iron pipes buried in the place where marine clay exists as described above, using statistical methods,
As a result of investigating the relationship between corrosion, environmental factors and burial period, the following facts were revealed.

(a) Marine clay showed strong corrosiveness, which contained a large amount of sulfur compounds and was characterized by strong acidification when oxidized.

(b) Marine clay tended to be more corrosive when coexisting with groundwater and when used for embankment.

(c) The distribution of the maximum corrosion depth of the buried pipe follows the log-normal distribution.

(d) When the relationship between the corrosion depth P and the burial period t is indicated by P = kt ⁿ , ⁿ should be calculated at a statistically significant level by considering k as a parameter indicating the corrosiveness of the environment. I was able to.

(e) In this case, by first removing the fluctuations due to environmental factors that account for a large percentage of the fluctuations in corrosion depth,
It was found that it becomes possible to reveal relatively small fluctuations such as the buried period.

(f) Further, the parameter n was constant without being present in the corrosiveness of the environment, and was n = 0.4 in the case of the cast iron pipe.

(g) From the above, the maximum corrosion depth of the buried pipe could be predicted at a practical level by several environmental factors and the burial period.

EFFECTS OF THE INVENTION As described above, according to the present invention, the growth rate of the pitting depth (P), which indicates the degree of external corrosion of the cast iron pipe buried in the soil, is calculated according to the power function of time (t): P = Kt ⁿ , and the parameter (k) is statistically calculated assuming that the parameter (k) depends on the environmental factors of the buried site, so the corrosion of the cast iron pipe can be predicted with high accuracy. can do.

[Brief description of drawings]

FIG. 1 is a flow chart showing a statistical method for predicting corrosion of a buried pipe based on the present invention, FIG. 2 is a view showing positions of respective variables in principal component analysis, and FIG. 3 is a maximum corrosion index on probability paper. FIG. 4 is a scatter diagram of measured values and predicted values of the corrosion index, and FIG. 5 is a scatter diagram of predicted values of the corrosion index and standardized residuals.

Claims (4)

[Claims] 1. A growth rate of a pit depth (P) indicating the degree of external corrosion of a cast iron pipe buried in soil is represented by P = kt ⁿ according to a power function of time (t), and the above ( k) is quantified by multiple regression analysis using attribute variables assuming that it depends on the environmental factors of the buried site, and then (n) is a linear model to the difference between the actual measured corrosion depth and the predicted value due to environmental factors. Corrosion prediction method for buried pipes, characterized in that it is obtained as a regression coefficient of. 2. Degree of external corrosion of cast iron pipe buried in soil
Corrosion index (Y₀) As the regression modelWhere i = 1,2, ..., m (number of samples) R: number of factors C_j: Number of categories in factor j δ_i(j ) = 1 (solid i reacts to category of factor j)
When) or 0 (at other times) a_i, a_{R + 1}: Coefficient e_iCorrosion of the buried pipe according to claim 1, characterized in that an error term is used.
Prediction method. 3. The parameter (n) is defined by the following equation: logY ₀ -logYc = C + nlogt where Y ₀ : Corrosion index indicating the degree of external corrosion Yc: Corrosion index based on environmental factors C: Linear represented by a constant The method for predicting corrosion of a buried pipe according to claim 1, wherein the method is obtained as a regression coefficient of a model. 4. A predicted value of a corrosion index showing the degree of external corrosion of a cast iron pipe buried in soil. Is Here, it is obtained by using Yc: a corrosion index based on an environmental factor, t: a burying period (years),
The method for predicting corrosion of a buried pipe according to any one of 1 to 3 above.